Social network X uses user data by default to train its AI, Grok.

Grok, the AI ​​tool created by Elon Musk, which arrived in France just three months ago, is already raising serious questions about its legitimacy on European soil. Friday, July 26, many users of the social network

In the privacy settings of the social network, activated by default and for all Internet user accounts, it is an option that allows you to transfer your data to xAI in order to train Grok, its generative artificial intelligence. To continually improve your experience, we may use your posts on X, your interactions with users, your Grok entries and the results associated with them for training and modification purposes.We read. This also means that your interactions, inputs and results may also be shared with the xAI service provider for these purposes. »

X returns in its parameters to One page, written in Englishwhich somewhat illustrates the features of this data sharing. The objectives of this training are: “Improving Grok’s understanding of human language and communication, improving Grok’s ability to deliver accurate, relevant, and engaging responses, developing Grok’s sense of humor and wit to make interactions more fun, and ensuring Grok remains politically neutral and delivers balanced responses.”

Read also | Elon Musk launches Grok, the equivalent of ChatGPT, on X

Disabled option

The same information page explains that X users have the possibility to refuse this processing, in two different ways: by making their account “private” or by unchecking the dedicated box in the settings (by selecting “Privacy and Security”, then “Grok”). The problem: this operation can only be performed via the browser, not via the X mobile application, on which the “Grok” parameter does not appear.

In the X social network settings, data sharing for Grok AI training is enabled by default.

Details that raise questions about the true capacity of L. Many Internet users were also surprised to receive no notice about the virtual activation of this data processing, which has been active since at least May, according to the research results. worldIn France, some claimed to have alerted the National Commission for Information Technologies and Liberties (CNIL).

Contacted, CNIL did not respond to requests made by worldX Corp, which has not communicated with the press since Elon Musk bought the social network Twitter in October 2022, automatically responded via email: “Currently busy, please come back later.” xAI, the company that developed Grok, also did not respond.

Read also | The material is reserved for our subscribers. CNIL: Surprising proposals from the Artificial Intelligence Commission to facilitate access to personal data

the world

Reuse this content

Leave a Reply

Your email address will not be published. Required fields are marked *